Machine Learning Techniques for Heart Disease Classification Using K-Nearest Neighbor Optimization With Particle Swarm Optimization Retno Wahyusari(a*), Eva Hertnacahyani Herraprastanti(b), Helmi Gunawan(c)
(a)Program Studi Informatika, Sekolah Tinggi Teknologi Ronggolawe
Jl. Kampus Ronggolawe Blok B No.1 Mentul Cepu
*retnowahyusari[at]gmail.com
(b,c) Program Studi Teknik Mesin, Sekolah Tinggi Teknologi Ronggolawe
Jl. Kampus Ronggolawe Blok B No.1 Mentul Cepu
Abstract
The disease is one of the causes of death. Based on data from the World Health Organization (WHO), there are at least 10 diseases that cause the highest number of deaths in Indonesia. Heart disease is ranked second after stroke. Cases of death due to heart or cardiovascular disease in Indonesia increased by 1.25% compared to the previous year. Considering the high death rate due to heart disease, accurate diagnosis is needed to prevent and treat heart disease. Diagnostic activities can utilize machine learning. Classification is a way of carrying out a grouping process based on certain characteristics for diagnosis. Some classification methods are decision trees, Naive Bayes, Support Vector Machine, and k-NN. k-NN is considered a simple method to apply in data analysis with many changing dimensions. Apart from having advantages, k-NN has a weakness, namely using all training data for the classification process which can result in a long classification process. Weaknesses of the k-NN algorithm can be overcome using the feature selection method. Particle Swarm Optimization (PSO) to solve optimization problems accompanied by various features. The results of heart disease classification using k-NN have an accuracy rate of 60.13%, while k-NN optimized with PSO has an accuracy rate of 90.75%, which is better than using k-NN alone.